0. E. BEEVOR, M.D., Lond., MR.C.P.,
ASSISTANT rilYSICIAN TO THE NATIONAL, HOSPITAL EOR
PARALYSED AND EPILEPTIC, LONDON.
iUEPlllNTEJ) FROM ^BRAIN,' Pari XXllI.']
PRINTED BY WILLIAM CLOWES AND SONS, Limited,
STAMFORD STREET AND CHARING CROSS.
•*w f' ■ » »» ,'r/r- -
C. E. BEEVOR, M.D., Loxd., M.R.C.P.,
ASSISTANT PHYSICIAN TO THE NATIONAL HOSPITAL POE
' PARALYSED AND ElILEPTIC, LONDON.
[UEPIUNTED FROM *BRAIN; Fart XZil/.]
PIHNTED BY WILLIAM CLOWES AND SONS, Limited,
STAMFORD STREET AND CHARING CROSS.
PRINTED BY WILLIAM CLOWES AND SONS, LIMITED,
STAMFOUD 6TKEET AND ClIARIN-a CKOSS.
THE CEEEBELLAR CORTEX.*
BY C. E. BEEVOK, M.D. BOND., M.R.C.P.
Assistant riiysician to the National Hospital for Paralysed and Epileptic, London.
Dr. Gaule, under whom I liave worked, and whom I have much
to thank for liis kind assistance, advised me to iise Wei^ert’s saure
fuchsia method (Centrhl. fur die med. ITm. 1882, p. 753), wliich
has produced excellent results. The first sections weie made
from pieces of cerebellum, one centim. square, hardened in three
per cent, soluiion of bichromate of potash and kept in a hot
chamber at 35° C. for 4-8 days. They were then washed in
water half an hour, then ])laced in alcohol 24 liours, and then
for 24-28 hours in a concentrated aqueous solution of siiuie
fuchsin, also at a temperature of 35° C. The pieces were then
slightly washed in water and put into .strong alcohol and dehy-
drated, imbedded in paraffin and cut into section.s, which were
then fixed to object slides, the paraffin dissolved (mt,'-* and the
sections washed out according to Weigert’s method with water
and alkaline alcohol.
In the second series of preparations they were further stained
with nigrosine after the above treatment ; and these preparations
proved the best. The third series were hardened in the same way,
but after fixing the sections on the slides, they were stained by
hajmatoxylin and eocene. This latter method of double staining
has alread}' been employed by Denissenko. In the fourth method
employed, the pieces of cerebellum were hardened by % chiomic
acid, imbedded unstained, and the sections fixed on the object
slide. The specimens were taken from the cerebellum of man
(adult and new-born) dog, rat, rabbit, hen and pigeon. My des-
cription is chiefly concerned with the dog, in which all the elements
are most clearly shown. The human cerebellar cortex is the most
developed, and as it is much richer in medullated fibres, and the
inedullated sheaths are better developed, it would undoubtedly
give better preparations, but unfortunately it is impossible to harden
the pieces directly after death and so obviate post-mortem changes.
And great stress is laid on the fact that in the preparations used
for this paper, in less ihan an hour after death, the pieces were
in the hardening fiuid in the warm chamber, which was used to
introduce a quicker diffusion in the substance of the piece.
’ This article is an extract from a paper published in the ‘ Archiv fiir Anatomie
und Physiologie’ (Physiologische Abtheilung) and from work done in Professor
Ludwig’s Physiological Institute in Leipzig.
* Canini, in these Archive, 188R, p. 155.
( 1 )
1. — Granule Layer.
Biifsis of desci'iplion, stinro fuchsin and nigrosine preparations.
In these the medullated fibres are seen, as described in Fig. 1 and
2 a, as a black or grej'- thread enclosed in the medullary sheath
stained red. It is very important to note that the axis cylinder is
dark and the medullated sheath red. If the saure fuchsin be not
sufficiently washed out, other fibres which are not medullated are
dyed red, but these have not the dark, central fibre which can be
stained by nigrosine or haematoxylin, and my description refers
only to these fibres.
Fig. 1 is a longitudinal section of a lobule of the cerebellar cortex
seen under a low power. Along the centre of this run the medul-
lated fibres, radiating out and passing through the granule layer
in the direction of the Purkinje cells. These fibres have been
described by Deiters,^ who maintained the connection of the
axis cylinder process of these cells with a medullated fibre.
Koschewnikofif'^ has isolated Purkinje’s cells in connection with a
medullated fibre. Hadlich^ says that from the white medullated-
fibre centre, fibres proceed to the gi’ey red granule layer, divide, and
pass as a priueipal nerve process into the lai-ge ganglion cell, into
every cell one fibre. Golgi ^ makes the axis cylitider process,
in its course thiough the granule layer, give otf branches at right
angles. Other descriptions are of a similar nature. But how very
indistinctly the fibres have been seen is shown by the very con-
tradictory statements, as for instance some say : that the axis-
cylinder runs direct to the medullary centre; that it runs in
a horizontal direction ; that this axis-cylinder soon gets a medul-
lated sheath, or that this occurs at a great distance from the cell;
and that the fibre is fine, — or thick. Nearly all maintain that it
divides. lladlich, and, later, Denissenko make reservations on
account of possible deceptions. Hadlich’s statements seem to me
the most correct.
Fig. 1 shows a clear pictm’e of the whole arrangement of the
nerve-fibres ; in each lobule some of them can be traced uninter-
ruptedly in their whole course to the immediate vicinity of the
Purkinje cells. The fibres leave the medullary centre with the
slightest change of direction. At the summit of the lobule they
radiate out like the out>tretched fingers of the hand. At the
lateral part of the lobule they ascend slantingly in a slight curve
through tlie granule layer, fi’hey run at definite distances from
each other, corresponding to the distances between the Purkinje
cells, to which their number, as far as can be ascertained, cor-
responds. At the bottom of the sulci the fibres are closely
aggregated and several pass through the gianule layer together.
' Deiters, ‘ Untersuclnuigen iiber Celiiru imd RUckeiimark.’ Herausgegebene
V. M. Sclmltze. 18(58.
* Koschewnikoff, ‘‘ Axencylinderfortsatz der Nerveuzellen iui Kleiuliirn dea
Kalhes;’’ M. Solmltze’s ‘ Ardiiv,’ 1869, Bd. v.
^ Hadlicl), “ Mittheilung iiber 'den Bau der menschliclieii Kleinhiruriude.”
Arcli. fiir luikr. Anat.,’ 1869, Bd. vi.
* Golgi, ‘ Centialbl. fiir die mod. Wiss.’ 1879.
( 5 )
Tlicpe fibres make a bend at Ihe border of the molecular layer,
lunning parallel with it for a short distance ^seo 1} j they do
not cross each other and do not branch ; but keeping the same
thickness, they run stiaight through the granule layer with
some slight wavings in their course. Their axis-cylinder is
thick, the medullated sheath slightly varicose, often presenting a
jointed appearance. In short, these fibres have in the granule
layer the same formation they had in the medullary centre. Each
remains an isolated medullated fibre w’hich, unbranched and in-
dependently, directs its w-ay towards a particular ganglion cell.
One can hardly doubt that tliese are their nerve fibres. Sections
are not very adapted to trace the passage of a fibre into a structure
of so much greater size. If the sections are very thin, the chances
of finding the right spot are extremely small, and this difficulty is
increased as the fibres always change their direction befiire enter-
ing the cell. Isolated preparations show more easily the connec-
tion, but then it is not certain that the fibres are of iho same
nature. I have therefore not been satisfied till I could confirm also
in section preparations the connection of the Purkiuje colls with
one of these fibres.
It can therefore now be said for certain that every Purkinje
ganglion cell is in connection with a distinct, isolated, medullated
nerve fibre. Besides the above described fibres, there are others,
which have already been seen and which caused the different
opinions concerning the fibres going to the ganglion cells, as it was
not known that there were two different kinds of fibres. Only
Hadlich ^ has as yet expressed this idea, and laid stress on the fact
that the many divisions described must refer to fibres of the second
kind. The second form of fibre produces a plexus which traverses
the gramde layer in all directions, passing into the medul-
lary centre on one side, and into the molecular layer on the other,
'fhese fibres branch about and anastomose, they arc of different
thickness, from the finest calibre up to that of the first kind of fibres,
their medullated sheath is less developed, is always varicose, often
interrupted, and sometimes the fuchsine reaction entirely fails. In
man these fibres ajipear far more numerous and better developed
than in the dog, when stained by saure fuchsin. Perhaps age
makes some difference.
It is quite clear that these fibres differ from those of the first
kind. The latter pass through them without being connected
vith them. Probably all branched fibres belong to the same
species and have the same function. I shall designate the direct
fibres of the first kind, joining ganglion cells as strairjht, or unhranched
fibres, those of the second kind as branched or anastomosing fibi es.
The spaces between the fibres in the granule layer are filled up
by the closely packed cells. Formerly these cells were all called
granules, and their cellular nature was doubted. Denissenko has
shown that we have here two kinds of cells, one stained by haema-
toxylin, the other by eocene. The haaniatoxylin cells correspond
' Hadlich, op. cit. p. 20.
( C )
to those formerly called granules, but they possess a miclens and
protoplasm, and therefore tlie attributes of a cell. Denissenko
regards these as connective tissue cells, whereas the eocene cells
are connected with nerve-fibres.^ The reason why this formation
was not recognised before, was because the cells are so tightly
packed. Very thin sections have to be stained by a dye which
will colour the protoplasm, for this is always present in the
form of a very small layer around a relatively large nucleus. The
protoplasm has numerous processes, very fine or thicker threads,
which form a thick network with the threads of other cells.
The threads refract light strongly and have a sharp contour,
the meshes are fine. They are best seen, where the cells lie less
packed together, at the borders of the granule-layer. Towards the
medullary centre there is no sharply-defined boundary, and
as in the granule layer there are everywhere fibres, so in the
medullary centre there are similar cells scattered everywhere.
But here the protoplasm is more abundant, the threads thicker,
and the meshes of the network are very much larger and join with
the narrow network inside the granule layer, re>embling the fine
framework in the white substance of the spinal cord. The nucleus
of these cells is very large and characteiistic. It possesses a dis-
tinct thread-like appearance, with the above-meTi tinned method of
hardening. The threads lie grouped in all kinds of forms in a
round clear space. If these nuclei were richer in colouring matter,
and if characteristic stages could be found, the idea of nucleus-
division might be entertained. Some of the nuclei are certainly
larger and their contour more distinct. The mass which is
coloured by hacmatoxylin is naturally that of the nucleus threads,
while the protoplasm, even when coloured by nigrosine, only
appears as a narrow line running round the nuclear spot. Between
the cells the medullated fibres make their way, and their medul-
lated sheath comes into most intimate contact with them. The
cell threads often seem to penetrate between the separate particles
of the medullated sheath. Denissenko describes these cells as
grouped round openings from which chmnels proceed, whose
walls are likewise lined by these cells. This appearance is seen
w'hen the cells lie particularly clos-e together. These channels are
only the medullated fibres which have penetiated between the
cells, whilst the place where the fibres cross each other gives rise
to the apparent openings. If the cells are less closely arranged,
the spaces occupied by these fibres are less compressed, and this
presents the appearance of cells in groups surrounded by fibres.
In these spaces other struefures are seen, which colour with
eocene and nigrosine, and called by Denissenko, eocene cells. They
seem to be interposed in the course of the medullated fibres, which
cress over the cells. They are often really interposed, as can be
seen by using nigrosine and saure fuch^in, when the cells
become the same colour as the axis cylinder which is then seen
* Henle and Merkel have already compured the granules to ihe cells of reticu-
lated connective tissue, and have tlierefore described them correctly according to
( 7 )
to be in connection with them, while the inedullated sheath is
'PJieir cellular character is doubtful, as I can find no nucleus
in most of them. The larger ones reiemblo multipolar ganglion
cells by their form and size, as also by their connection with
nerve fibres ; and as it is often difficult to discover a nucleus in
ganglion cells when the protoplasm is colourerl, these are probably
ganglion cells. The smaller of these cells are often deficient in any
cellular character and appear to be thickenings along the axis
I think I can with certainty explain the use of the limmatoxylin
cells, viz. that of glia-cells. These have not only nuclei, but are
cells with protoplasm and processes; the processes form a network
which extends through the whole granule layer, and which is
also connected with a similar network in the medullary centre.
These appearances are just the ciiteiion of glia-cells, and if one
accepts glia-cells at all, they must bo cells like those above
described. Schwalbe has lately objected to the glia-cells generally,
from the hypothesis that the cells can be separated from the fibres
of the finest network. But these objections are mostly theoretical.
Besides, in my preparations, the network of the fibres proceeds
from the body of the cell (see plate Fig. 2, h) and this agrees with
what Henle and Merkel give in their work on the supporting
structure of the central nervous system. Also in the description
of the supporting framework of the molecular layer I have only
to confirm their observations, agreeing with them that these
cells are glia-cells, but doubting that they are cells of reticulative
connective tissue, or that they may become ganglion cells. Since
the embryologists have shown that the central nervous system
proceeds from the ectoderm, and that there exists a thorough
separation between the epithelial and connective tissue structures,
the “ archiblastic and parablastic ” formations, it is probable from
embryological researches, and what we know of its chemical
nature, that the glia is an epithelial formation.
The glia stands in relation to the ganglion cells and their
protoplasmic processes, as the inedullated sheath does to the axis
cylinder of the peripheral fibres. It is a continuation of this latter-,
and there are forms of transition. The cellular elements are
common to both ; from this can be explained, what might be con-
sidered an objection to the definition of these granule cells as
glia-cells, viz., their accumulation into distinct regular layers.
Where numerous fibres lose their medullated sheath (those of the
first type), or acquire it (those of the second typo,) there are to be
lound accumulated the cellular elements which are common to the
glia and the medullated sheath, and where this occurs in regular-
planes in the cerebellum, the position of the cells is found to
2. — Molecui.ar Layer.
a. The ground Substance . — This layer owes its name to the
peculiar substance forming its foundation. The older authors
( 8 )
describe it as a finely granulated mass, in wliich a finer structure
is not recognisable; however Fromman has already shown it to
be a sponge substance built out of a network of threads. Kiihne
and Ewald haA’’e proved that we have here a formation of a
peculiar chemical structure, analogous to the horny framework
of the medullated sheath of the peripheral nerves. Schwalbe
describes the suV>stance in his ‘Lehrbuch der Neurologic’ as a
fine network, the fine trabeculee being composed of neuro-keratin e.
This opinion is also borne out by my plates, Fig. 2, b-d. The clear
substance lying in the enclosed spaces of this network becomes,
with the means used to clear up the medullated substance, as
clear as this latter ; the osmic-acid luethod cannot be applied, as
the distinction between the trabeculje and its meshes only comes
into view with very thin sections, and these can only be made
after treatment with alcohol. The saure fuchsin method, however,
produces a very characteristic action on erythrophile substances.
In many of the mesh-spaces, but not in all, a fine red line bounds
the clear contents, in the same w^ay as seen in the cross-section
These red-bounded spaces are larger or smaller in size, and some-
times several in a row. But they are not sections of fibres, as can be
seen from the absence of any axis c}dinder ; neither are there,
where they lie, so many fibres visible, which is easily proved by
alteiing the direction of the section. It thus becomes probable
that as this network corresponds to the horny framework of the
medullated sheath, so also this intermediate substance corresponds
to the myelin ; but not chemically, for even the medulla substance
of the periphery differs fi om that of the central nervous system in
not re-acting to saui’e fuchsin ; and we must consider that the relation
of the neuroglia of the central nervous system diflfers more from
the peripheral nerves than does the medulla of the fibres of the
central nervous system. There are, however, many transitions, viz.,
the above described varicose sheaths of the brancihed fibres, which
often form separate round formations and the glia- nets containing
erythrophile substance. Another analogy becomes visible, when
we consider the cellular elements of both, viz., the glia and the
medullated sheath. Eanvier has acquainted us with the cells of
the medullated sheath, and he concludes that the myelin is enclosed
in the meshes of their protoplasm. Tow'ards the sheath of
Schwann and the axis-cylinder the protoplasm is thickened to form
the inner and outer protoplasmic sheath. When this theory was
proposed, Kiihne and Ewald had not yet discovered their horny
sheath, which probably corresponds to Kanvier’s protoplasmic
sheath, if neuro-keratine be substituted for pn>toplasm. I think
it probable that from the protoplasm of the yimng cell, neuro-
keratine is formed by differentiation, as is similarly described by
Waldeyer in the horny change of the epideimis cells. There, a
stroma of keratine threads foims itself out of the protoplasm of the
cell, while in its meshes, a fatty substance, eleidin, is deposited.
It is difficult to distinguish between the sheath and the framework.
Either can resolved into neuro-keratine. Probably transitions
( 9 )
between the two will be found. I have also tried to convince
my.self in preparations of the horny sheath, treated by Kuhne’s
method and coloured by baematoxylin and nigrosine, whether the
neuro-keratine threads really are in connection with the nuclei of
the medullated sheaths, and I can assert that one can really see
this, so far as it can be observed. Assuming the horny framework
of the peripheral nervous system to be a formation of cells, its
relation to that of the central nervous system appears all the more
assured. For not only the network of the granule layer already
desci ibed, but also the network of the molecular layer, appears as
glia-cells. The connection between both is best observed in the
collection of glia-cells to be found at the base of the molecular
layer. \\ hen from any reason, the molecular and granule laye’S
are separated — which may be caused even by pressure of the
cover-glass — there remains one or more layers of cells separated
with the molecular layer. The separation occurs below the
Pnikinje cells, where the unbranched fibres bend round and
spread out. Here, the layer of glia-cells attached to the mulecnJar
layer forms a membrane, by joining their processes, to support both
the fibres and blood-vessels spread out under it, and the Purkinje
cells resting on it, and to form also the groundwork of the
molecular substance. In Fig. 2 fc, this is shown in section. Ils
cells are generally larger than those of the granule layer, as are
also their protoplasm and processes, which here become very
distinct. Many are pyramidal or pear-shaped, with their broader
basis turned to the granule layer, and form, with the long processes
given otf from the glia-nets of the granule layer, that mem-
brane of the molecular layer which may be represented as a
fiat network broken up by the cells interspersed in its ineslies.
Through this membrane pass all the numerous fornialions on their
way from the granule to the molecular layer, and vice vensa. As
the name of limitana externa has been adopted by Henle and Mei-kcl
for the layer of the glia-nets which is in contact with the pia mater,
and this layer, by its formation and origin being completely
analogous to the above-mentioned membrane on the inner surface
of the molecular layer, I propose to call the latter limitana interna.
Some of the processes from its cells, which are vertical to tlje
limitans interna and go off from the apex of the cells, are parti-
cularly thick and can be traced far into the glia-network; and at
regular intervals these processes are still thicker and reach to the
limitans externa, to which they are attached by a flattened pedicle
and serve as supporting pillars to the molecular layei-. These
were discovered by Bergmann and described by Ilenle and Merkel.
They have also described the limitans externa as a special mem-
brane, separable from the pia mater, and consisting of glia threads.
Obersteiner has described one or two c 11 layers under the pia
mater in the new-born animal, from which layers is developed the
limitans externa, as well as the supporting pillars; probably the
limitans externa has the same character in the new-born as the
Innitans interna in the adult, so tliat the pedicles are remains of
pyramidal cells, whose apex processes formed supporting pillars
( 10 )
ami basal processes the network of the liinitans externa. 'J’ho
network between the membrane is also piobably the remains of
the cells described by Obersteiner, as filling up the embryonal
molecular layer, and which have been absorbed with the growth of
the network ; as the network is too extended to have been formed
from the supporting fibres, or from the glia-cells scattered in the
molecular layer. Perhaps the protoplasm of the embryonal cells
may have been converted info neuro-keratine. Two further points
about the glia-cells remain to be discussed. (1) Their relation to
the ganglion cells and their processes; (2) their relation to the
(1) We have already mentioned that the glia-cells are accumu-
lated round the ganglion cells (see Fig. 2 b). Whilst the ganglion
cells at their lower part rest on the limitans interna, they are
packed at the sides by these heaped up glia-cells, so that the cells,
together with the limiting membrane, closely surround a hollow
in which the ganglion cell lies. At the upper and lower part the
processes pass out. This walling-in is seen best when the Purkinje
cell has been detached from its space, which is then seen spun over
with a network of fine fibres, which are continued at the sides into
the bodies of the glia-cells. This capsule is thus formed by a net
of neuro-keratine threads, supported by a denser accumulation of
cells similar to that found in the limitans interna. A somewhat
similar capsule has been described by Denissenko and by former
authors. Of course such a capsule consisting of a network will
allow the cell a free connection on all sides ; there is, however,
a special condition for the axis cylinder and the protoplasmic
processes. The former seems to acquire a continuation of this
capsule as a sheath, which is connected with the medullated sheath
of the nerve-fibre. But I have no definite drawings on this point.
As is evident from Fig. 2, b, c, everywhere in the course of the
protoplasmic processes- the glia-nets suriounding them ‘become
thickened, so that numerous thieads seem to be attached to them,
whilst on each side, the meshes of the glia-nets become larger.
This attachment of glia-threads has been described by Hadlich
and other authors, who state that the protoplasmic processes appear
rougli, or that extremely fine threads go off from them laterally.
This attachment is however only apparent ; it is in reality only
a tube-like thickening of the network around these processes,
forming a sheath of neuro-keratine threads, continuous with the
capsule of the ganglion cells. This appeaiance resembles the axis-
cylinder of the peripheral nerve fibre (when treated according to
K-uhne’s method), which is closely surrounded by a thickening of
the horny structure, decribed by Kiihne as the inner homy sheath,
and here the keratine threads often seem to join the axis-cylinder
itself. The broad meshes of the glia-nets on the sides of the
protoplasm processes correspond to the space between the inner
and outer horny sheaths, which is also traversed by a few threads.
A formation analogous to the outer horny sheath is absent in the
glia. The resemblance between the relation of the protoplasmic
process to its glia, and the relation of the axis-c}’linder to its horny
( 11 )
sheath helps very much the comprehension of the central nervous
system. But the sheaths in the glia are not isolated, they are in
connection everywhere, and are only interrupted by the branching
out of the processes. In the glia, the homy component part is
unduly in excess of the thread network, as compared with the fatty
part enclosed in the meshes : and the fatty substance is not identi-
cal with mjmlin of the peripheral nerves. As Weigert’s reaction
is only sometimes shown, probably different gradations of myeline
exist in the nerve-fibres of the periphery and in those of the central
nervous system, where varicose fibres occur, increasing as they
(2) It is not easy to establish the boundary between the glia and
the connective tissue. When they were still unconditionally classed
together, it was observed that there by no means existed an
immediate connection between the pia and the glia lying under it.
Ilenle and Merkel — who in their treatise assume that glia, nerve
elements and connective tissue are produced out of the same materials
— describe, however, that the glia limits itself by a lamina limitans
externa, and that this is easily detached from the pia, being
separated by spaces traversed by single threads, — subarachnoid
lymph spaces. A retraction of the glia from the connective tissue
sheaths of the vessels leads to the production of the peri-vascular
spaces. The glia therefore is only in connection at single points
with the connective tissue, which is necessary to give support to the
whole framework. In short there seem to be connecting-fibres
only between the pia and the limitans externa, as these fibres do not
sink deeper into the glia-network.
I'he connective tissue of the pia may be compared with the
Schwann sheath of the peripheral nerves, if the glia be ounpared
with the mednllated sheath ; and as between the Schwann primitive
sheath, and the medullated sheath, so between the connective tissue
and the glia, there exists only a loose connection ; from this it does not
follow that the glia might not histologically bo reckoned with con-
nective tissue. But this depends entirely on whether or no the glia-
cells are descended from epithelium cells of the ectoderm, for our
present classification of the tissue is essentially generic. From the
above, it is no lonu.er possible to treat of the glia as an undefined
boundary between nervous and connective tissue. We have then in
the nervous system three distinct systems lying close together :
1. Axis cylinders. Ganglion cells. Protoplasm processes.
Medullated sheaths. Glia-cells. Glia net-works.
3. Schwann’s sheath. Pia mater. Connective tissue-sheath.
Of these systems the first only can be designated as nervous.
h. The Nervous elements . — One link in the chain of this nervous
system fails us. The branched processes of the Purkinje cells
must enter again into connection with nerve fibres. It is easy
to trace the branches to close to the limitans externa, but here
every diiect trace is lost. Kindfleisch and Strieker suppose them
to become lost in the ground-substance of the neuroglia, but
( 32 )
tin’s has been very much contested, as it is against the physio-
logical demand for a definite course. In my preparations I find no
indication of a connection of the ganglion-cell processes with the
threads of the glia network, A second view makes the pro-
cesses end in peculiar nuclei, which lie at the periphery of the
molecular layer. I have found it difficult to discover what form
of cells Denissenko means by these peripheral nuclei, for at
the limitans externa or towards the inner part of the periphery,
there are very few cells to be found in the adult animal. Nu-
merous cells are present in the young animal onlj*. Denissenko
calls these cells nuclei, and yet states that they are very little
coloured by nuclei staining fluids, and further, that in some animals,
fibres spring from them which penetrate vertically the molecular
layer. The latter fibres correspond to the supporting fibres of the
glia, and I consider that the so-called peripheral nuclei are the
pedicles with which the supporting fibres attach themselves to
the limitans externa. The degree of coloration mentioned by
Denissenko agrees also with these pedicles, which might also be
called nuclei, as they are possibly the remains of cells; but with
the processes of thePurkinje cells these pedicles are not connected.
This view of Denissenko agrees with Golgi’s, avIio makes the
processes of the ganglion cells end in connective ti.-sue corpuscles
at the periphery, which Denissenko even conjectures to be partly
identical with his peripheral nuclei. With the excei3tion of the
sheaths of the vessels, I must decidedly dispute the presence of
real connective tissue corpuscles at the periphery of the molecular
substance on the inner side of the limitans externa. A third
view makes the processes of the periphery bend round and run
back to the granule layer ; and to prove this, certain fine straight
fibres, which traverse the molecular layer vertically to the
periphery, have been pointed out as these processes (Boll) ; but
the greater number of these fibres can be traced to the limitans
externa where they end in pedicles. They are the supporting fibres,
as others have already showm.
Lastly, Hadlich has suggested that the ganglion cell processes
return again in the same form. This appears at first difficult to
contradict, as the processes ma^^be either running out or returning.
But Hadlich has already conjectured that there might be two kinds
of nerve-fibres, one entering into connection with the axis cylitider
of Purkinje’s cells, the other with the already mentioned returning
branched processes. The latter therefore would not return to gan-
glion cells, but join directly with nerve fibres. In this way we have
a distinct criterion by which to recognise the out-going from the in-
going fibres, and a proof of whether they really exist. I have looked
through my preparations and found that there are numerous places
where, in the deeper par ts of the molecular layer, all ganglion cell
processes can be traced to their corresponding Pui kinje cells. 'J’his
precludes the possibility that the protoplasmic processes, after
spreading themselves out at the periphery, should collect together
and return in the same form to the granule layer, to join with nerve-
fibres. Another hypothesis assumes that the reluming processes
( - 13 )
reassemble in the ganglion cells, of which there are two kinds, and
which are not to bo distinguished in tbeir outward appearance. But
this is contradicted by the fact that the branched mednllated fibres
can be traced far into the molecular layer. The object of these
fibres could not be seen, if the protoplasmic processes return to a
o-anglion cell lying at the border of the granule layer before join-
Tiig the neive fibres. Besides tliis the number of unbranched
fibres is probably just as largo as that of the ganglion cells, so that
we can drop the hypotho'is of two kinds of ganglion cells. If the
number of the unbranched fibres is just as many as the Purkinje
cells, and if they end at the granule layer, whilst the branched
fibre’s extend into the molecular layer, the probability becomes very
great that this cell is in connection with both kinds of fibres, with
those in the granule layer by its axis cylinder process, with the
others in the molecular by its branched protoplasmic processes.
Fi<>-. 3 is drawn from a section cut parallel to the surface, and
therefore at right angles to the direction of sections hitherto shown.
1’he drawing represents the bottom of a sulcus, so that not only the
molecular layer, but also the Purkinje cells and a portion of the
granule layer are visible on each side. The medullated fibres
appear on the flat surface of the section, running in two directions
at right angles to each other, and round the Purkinje cells are
numerous fibres passing into the molecular layer. The section here
given is very thin, or else the fibres would be very numerous, as is
the case in thick sections, where it is difficult to see individual fibres.
This horizontal plexus is most dense in the deeper parts of tlie
molecular layer, around and immediately above the Purkinje cells.
But also in the more superficial planes, fibres running in the same
direction can be seen, and the better the reaction has succeeded,
the further does their region extend. They are all much finer than
the unbranched fibres, but of the same calibre as the finer branched
fibres. In the corresponding vertical sections of the lobule, fibres
are seen in bundles close round the capsule of the Purkinje cells
and passing into the molecular layer. Little of the above described
horizontal plexus can be seen in these vertical sections, as the one
set of fibres is cut across (compare Fig. 2, 6), while the other runs
parallel to the surface. On the other hand there are fibres visible,
Avhich descend vertically in the third plane. These are also fine
and often run close to the supporting fibres. Generally these
cannot be traced further than through half the molecular layer.
Beyond that, although ei ythiophile substance is met with, it is not
arranged in the form of fibre sheaths, but in isolated rings, without
any connection. Probably these single rings must stand in some
relation to the continuation of the system of medullated fibres.
As the connection of the fibres of the molecular layer with the
fibres of the second kind in the granule layer is very distinct, we
may say that there exists here a fibre system, which, starting as a
plexus from the medullary centre, branches through the granule
layer, until the finest fibres are lost in the molecular layer, whilst
their medullated substance can be traced further than the axis-
( n )
As regards the connection of tlie separate elements it is easy to
establish the following points of a scheme :
1. Each unhmnchecl fibre is connected with each Puricinje cell. The
axis-cylinder passes into the protoplasm of the cell, the medullated sheath
into the glia-like capsule of the cell.
2. The axis-cylinder becomes converted into a number of fibrils in the
cell, which pass into the branched protoplasmic process, the fibrils run
in the process, ivhich is surrounded by a glia-sheath, as completely distinct
threads as far as the periphery. In the branching of this process, the
numerous fibrils, contained in it ichere it leaves the cell, are gradually
distributed until they become isolated.
My preparations show that the arrangement of the ganglion cells
corresponds to M. Schultze’s scheme. Obersteiner has already
described this for the Purkinje cells. Sometimes fibrils cross each
other at the points where the processes branch.
3. The fibrils thus isolated bend round at an angle of 90° (not 180°),
spread themselves out in a plane lying parallel to the surface, and re-
arranging themselves in definite order as fibres surrounded by a medullated
sheath, run, with frequent interchange between the fibres, in the form of a
plexus back to the medullary centre.
This third point of the scheme contains, of course, much that is
hypothetical. But there can be litile doubt that the fibres of the
second system serve to establish somehow a connection with the
branched processes of the ganglion cells. To explain why the
fibres should not bend at an angle of 180°, instead of 90°, I must
refer to Obersteiner’s discovery that the protoplasmic processes of
the Purkinje cells only spread out in a plane at right angles to the
surface, and also to the direction of the lobule and the plane of the
medullated fibre laj^er. On the other hand, the medullated fibres
of the molecular layer run in planes parallel to the surface, and at
right angles to the planes of the processes. The processes only
need now to bend round at an angle of 90° to pass into the plane
of the fibres. Certainly other fibres are required which shall turn
at an angle of 90° and connect the higher lying planes with the
lower, but these fibres would be coarser and medullated. The
strictly angular arrangement of the elements of the cerebellum is
remarkable, both in the nervous elements and in the supporting
structure. The important point is that we do not see the con-
necting elements between the ends of the protoplasmic processes
and the beginning of the nerve fibres ; and in the absence of any
really decisive result, some other solution of the problem is possible.
Perhaps future research will reveal that just at that point where
the continuity of the elements is lost, an arrangement intervenes
which we cannot now recognise.
Further, there are medullated fibres under the pia mater Ij ing
on thelimitans externa, broad fibres of a similar character to those
of the medullary centre.
The same fibres may be seen passing through the molecular layer
in a slanting direction. A second collection of medullated fibres of
the same character lies half-way up the molecular substance.
Both imn parallel to the medullary centre. As those fibres are
( 15 )
only found in a KUiall minority of the lobules, lliey are pr.-bably
an aberration of the medullary centre.
Sometimes the Purkinje cells of the same lobule differ in ap-
pearance. Some show, as seen in Fig. 2 6, a distinct nucleus,
with the nucleolar contents surrounded b}- a ring of nuclear
substance, and in the protoplasm, a distinct fibrillous formation
which completely fills up the glia-capsule and in the double-
staining with nigrosine and saure fuchsin are stained a light grey
colour by the foimer only.
In the other kind a nucleus is hardly to be distinguished, and
no fibril-formation ; the cell appears as a homogeneous body which
remains dyed by same fuchsin, even after washing out by alkaline
alcohol. The cells and their processes are slightly retracted from
the glia, and the relation between the capsule and its sheath is not
quite so distinct. It is not probable that there arc here two kinds
of ganglion cells, as the difference in tlie cells is only seen in some
of the lobules. It is hardly due either to different action of the
hardening re-agent or to the fact of one cell dying before the other,
as an external influence would act differently on the cells of two
distant lobules, but not on those of the same lobule. We must
conclude that the condition of cells must haVe originally been
different at the moment of death.
EXPLANATION OF THE PLATES.
o Unbranchcd broad nerve-fibres of the first kind.
a' Purldnje’s cells.
a" Their protoplnsmic processes.
o'" Branched fibres of the second kind.
/3 Glia-cells of the granule layer (granules, glia-cells).
Cells of the membraua limitaus interna.
Supporting fibres (long radiating processes of the glia-cclls).
0"' Glia network (short branched processes of the glia-cells).
Membrana limitans externa.
7 Pia Mater.
5 Non-nucleated structures of the granule layer, stained by Eocene and Nigrosin
GENERAL "VIEW MAGNIFIED WITH A LOW POWER.
Fig. 1. Longitudinal section of a lobule from the cerebellar cortex. Medullary
centre— internal or granule layer — external or molecular layer. The pia mater
with its blood-vessels is slightly separated. Shows the radiation of tl;e un-
brancliecl fibres from tlie medullary -centre tlirough the gmiiule layer to tie
Fig. 2. Details of the same section seen by a higher power U. oil immersion)
show ing ; ^
a. Granule layer.
h. Boundary between granule and molecular layers.
c. liliddlo ot the molecular layer.
d. Outer part of the molecular layer.
The figs. 2 a-d are to be taken in series.
2. a, allows fibres of the first kind, a, and of the second kind a'" between
the glia-cells, |8, and the eocene cells, 5.
2 b. A Purkinje’s cell, a' with its protoplasmic process, a" which breaks
up into fibrils. The glia-cells of the lamina interna, )3', some of them large
and rich in protoplasm, are heaped around the Purkinje’s cell. The glia net-
work )3"' proceeds from the processes of the glia-cells, some supporting fibres
/3" are here not sien joining the cells. In the meshes of the glia-nets are seen
sections of medulTated fibres of the second kind, a'", stained red.
2 6‘. The glia-cells, in the interval between two Purkinje’s cells, form by their
intimate connection the limitans interna, )3'. Some pyramidal-shaped glia-cells
are elongated into processes which extend direct to the periphery and form
supporting fibres j3".
2 6*. A Pm-kinje’s cell in the glia capsule. The cell is seen, so as to show
clearly its capsule. The threads of the glia are partly visible coursing over the
2 c. Glia network in the middle of the molecular layer. Very clear broad
trabeculae of the network, /3"', which is traversed by parallel supporting-fibres
i3", and also by a branching process from a Purkinje’s cell, showing fibrils, a".
2 cb A supporting fibre, ;8", in connection with one of the solitary glia-cells
found in this region.
2 d. The periphery of the lobule, with the limitans externa )3'" cut across at
right angles, and with the supporting fibres, /3", attached to it by broad pedicles ;
between them is the glia network.
Fig. 3. Horizontal section (parallel to the surface) through a lobule of the
cerebellar cortex near the boundaiy between the molecular and granule layers,
' which is depicted at the periphery of the drawing, so that the Purkinje’s cells
appear surrounded by medullated fibres of the second kind which having reached
the molecular layer, spread out in the plane of the section, in two directions at
right angles to each other.
These plates are faithful representations of the preparations, and are not in
any way schematic.
LONDON: I’lUNTED DY WILLIAM CLOWES A'D SONS, LIMITED, STAUt'ORD STKEKl
AND CHARING CROSS.
».v f • •